Landon 2005 (18)


Prospective cohort study. Data collected for years 2000-2005

HIV patients age 18+ in federally funded HIV clinics (community health centers, community-based organisations, health departments, hospital outpatient clinics, university medical centers) (n = 5247).

ID physicians vs. (HIV-expert) GM physicians vs. (non-HIV-expert) GM physicians

ID physician



After controlling for specialty training, caseload and knowledge, non-HIV-expert GM’s overall quality of care is lower than ID physicians and HIV-expert GM (adj P = 0.02). Caseload differences not statistically significant. All data were adjusted for patients’ characteristics.

cART therapy

ID physicians and HIV-expert GM vs. non-HIV-expert GM, 80% vs. 73% (P < 0.001)

VL control

ID physicians (41%) vs. HIV-expert GM (39%) vs. non-HIV-expert GM (31%) (P = 0.01)

Flu vaccination rate

ID physicians (54%) vs. HIV-expert GM (49%) vs. non-HIV-expert GM (41%) (P < 0.01)

PJP, PPD, HCV, Pap smear

No statistical significant difference

Outpatient visit

ID physicians (66%) vs. HIV-expert generalists (69%) vs. non-HIV-expert generalists (57%) (P < 0.01)

Page 2003 (17)

Zurich, Switzerland

Prospective cohort study. Data collected for years 1999-2001

HIV patients on ART receiving care in an urban area in general practices vs. a specialized university o utpatient clinic (n = 106)

GP vs. ID specialists vs. shared care between GP and ID specialists.



Shared care


*= Data measured at baseline only. No adjustment was made because patients in all healthcare provider groups had similar baseline characteristics.

Median CD4 count

459 cells/µl (range: 120 - 1442)

449 cells/µl (79 - 891)

303.5 cells/µl (104 - 751)


% of patients with VL < 400 copies/ml





Mean quality of life of all participants

129.5 ± 22.5 baseline, 131.6 ± 25.6 at study end


Mean patient satisfaction*

Patients with GP and shared care are more satisfied than those with ID specialists

P < 0.05

ART use

No statistical significant difference

Patient’s evaluation of providers

GP received statistically significantly higher mean scores for many dimensions of care (such as competence, information, communication, time and trust) than ID specialists.

Mean rate of ART adherence

No statistical significant difference

Schneider 2008 (19)


Prospective cohort study. Data collected for years 2001-2003

HIV patients admitted to a general medicine, non-critical care inpatient services in six institutions (n = 1207).

Hospitalists (spend at least 25% of time in GM services) vs. non-hospitalists (spend less than 25% of time in GM services).



Readmission rates, emergency department visits and health status were self-reported. No clear descriptions of hospitalists and non-hospitalists besides the amount of time they spend in GM inpatient care. All data were adjusted for patients’ characteristics, physicians’ experience, care sites and time of hospital admission.

In-hospital mortality rate

adj P = 0.38

6-month mortality rate

adj P = 0.57

30-day readmission rate

adj P = 0.99

30-day emergency department visit rate

adj P = 0.70

30-day reported health

adj P = 0.97

Overall patient satisfaction

adj P = 0.98